AI capacity is the compute, power, cooling, and networking infrastructure needed to run artificial intelligence workloads. It includes GPUs or other accelerators, the servers that host them, the electrical feed to support high-density hardware, and the network paths that move large data sets quickly enough for training and inference.
It matters in cyber security because scarce AI capacity can become a bottleneck during modernization and defense. Security teams may need it for malware analysis, anomaly detection, log correlation, and automated response, while attackers may use the same resources for large-scale model training, phishing generation, or password-guessing at speed. When capacity is limited, organizations may postpone controls, overload existing systems, or create outages during migration. Defenders plan around this by sizing infrastructure carefully, isolating critical workloads, and monitoring power, cooling, and network usage as part of operational risk.